Model-Based Development (MBD) is posed as an integrated end-to-end simulation-driven and code-generation-based process for developing complex systems that are composed of sensors, actuators, software, and functional payloads. Spacecraft and Launch Vehicles provide good examples of such systems, therefore these systems are featured in this paper to illustrate the range of functionality and complexity that can be addressed using MBD. The examples shown herein exploit block-diagram programming to simulate systems, integrate legacy code, connect to device drivers and generate code targeted to pre-existing operating-system-agnostic middleware or directly to embedded processor operating systems. The effectiveness of using a single simulation as the source for auto-generated embedded flight software as well as real-time plant simulations for use in processor-in-loop and hardware-in-loop test systems is characterized. The motivation for adopting end-To-end MBD is to provide substantial reductions in system development time and cost while enhancing the robustness and reliability of the subject system. The paper provides insights into these metrics in the context of low-cost navigation, guidance, and control avionics for spacecraft and small launch vehicles. Clearly, the effectiveness of this process is dominated by the fidelity of the underlying dynamic system models, so the paper provides examples of detailed subsystem modeling including effects of uncertainties and sensor measurement errors and addresses real-time simulation issues. Since the MBD process creates a system design that reflects the behavior and fidelity of its underlying simulation, realization of the design as hardware and embedded software must be tested in “real world” environments, so the paper cites examples of how data from hardware tests supports evolution of simulation fidelity as well as embedded software robustness and reliability. In MBD, testing is accomplished via simulations and PIL/HIL test systems initially and then augmented by hardware testing to obtain excitation-driven data that enhances simulation model fidelity and prediction accuracy.